Undiagnosed diabetes and impaired glucose regulation in adult Ghanaians using the ADA and WHO diagnostic criteria.

Fasting glucose and oral glucose tolerance test (OGTT) criteria for glucose homeostasis were compared in a cross-sectional cluster, community study in Accra, Ghana. A total of 4636 subjects without prior diagnosis of diabetes had fasting plasma glucose, 2-hour OGTT and measurement of cardiovascular risk factors. Mean age of subjects was 44.2 years: 39.1% of subjects were males. The overall prevalence of undiagnosed diabetes ascertained with both criteria was 4.5% (n=209). The prevalence of undiagnosed diabetes by fasting (3.2%) and OGTT (3.1%) criteria were similar (p>0.05). The prevalence of impaired glucose tolerance (IGT) (15.8%) was higher than that of impaired fasting glucose (IFG) (10.7%). Only 56.5% (n=83) of subjects with diabetes by fasting criteria also had diabetes by OGTT criteria. Sixty-two subjects (42.8%) with diabetes by OGTT had normal or impaired fasting glucose. There was poor agreement between the two diagnostic criteria (kappa=0.31). The concordant normoglycaemic group was the youngest and had the lowest body-mass index (BMI), waist girth, waist-hip ratio (WHR), total cholesterol, and systolic and diastolic blood pressures. The concordant diabetic group, in contrast, had the highest BMI, waist girth, WHR, total cholesterol and triglyceride levels. Both systems gave similar undiagnosed diabetes rates but dissimilar IFG and IGT rates. There was poor agreement between the two diagnostic criteria. Diagnostic criteria influenced cardiovascular risk factors. A case may be made for using both criteria in order to ascertain all "diabetes" and all "at-risk" subjects.